How we measured
20 companies, each scanned in September 2026 against ChatGPT with web search, Gemini, Google AI Overviews, Perplexity. For each company we read its website, wrote about fifteen buyer questions from what it sells, put every question to every engine and recorded who each answer named, in what order, and which websites it cited. That is 1,014 engine answers. Scores are the same ones the report uses: visibility from the answers, readiness from thirteen technical checks on the site, and an overall that blends them three to one.
What this does not measure: market share, revenue, or whether the engines are right. It measures what a buyer is told.
Who the engines name instead
When one of the 20 companies was missing from an answer to its own buyer question, this is who was named in its place, counted per question lost. Dynatrace took 48, more than the next four combined (52).
| Named instead | Questions taken |
|---|---|
| Dynatrace | 48 |
| New Relic | 15 |
| Rootly | 13 |
| Datadog | 13 |
| PagerDuty | 11 |
| ServiceNow | 5 |
| BigPanda | 4 |
| AppDynamics | 4 |
Two clusters sit inside the table. The monitoring and observability questions go to Dynatrace, New Relic and Datadog. The incident-response questions go to Rootly, PagerDuty and incident.io, and a company that sells both is losing to two different sets of names.
The 20 companies, scored
| Company | Category, as its site reads | Overall | Visibility | Site readiness |
|---|---|---|---|---|
| Atlan | enterprise AI context layer | 70 | 65 | 85 |
| SolarWinds | IT monitoring and observability software | 60 | 52 | 85 |
| BigPanda | IT operations software | 57 | 48 | 85 |
| ClickHouse | open-source column-oriented database | 57 | 52 | 73 |
| Resolve AI | AI incident response software | 49 | 34 | 94 |
| Kentik | network monitoring software | 48 | 46 | 54 |
| Bigeye | enterprise data observability and AI governance software | 46 | 34 | 83 |
| ScienceLogic | AIOps platform | 46 | 35 | 77 |
| OpenObserve | observability platform | 44 | 30 | 85 |
| Metoro | Kubernetes observability software | 43 | 30 | 83 |
| Monte Carlo | data and AI observability platform | 42 | 30 | 77 |
| Fabrix.ai | agentic AI platform for IT operations | 41 | 23 | 94 |
| Checkmk | IT monitoring software | 38 | 22 | 87 |
| Selector | network observability software | 36 | 23 | 77 |
| Sherlocks AI | AI SRE platform | 36 | 14 | 100 |
| Virtana | hybrid observability platform | 34 | 19 | 77 |
| Causely | autonomous service reliability platform | 32 | 15 | 85 |
| Traversal | AI site reliability engineering platform | 28 | 18 | 58 |
| Netdata | infrastructure monitoring software | 27 | 19 | 52 |
| Middleware | observability software | 24 | 13 | 58 |
Median overall 42, median visibility 30, median site readiness 83. The category column is what each company's own website says it is, which is what the questions were derived from.
Which sources the engines cite
Across every answer, the domains cited most often as the source for what was said:
| Source | Answers citing it |
|---|---|
| gartner.com | 185 |
| newrelic.com | 148 |
| youtube.com | 121 |
| datadoghq.com | 114 |
| dynatrace.com | 107 |
| rootly.com | 87 |
| augmentcode.com | 69 |
| g2.com | 68 |
| incident.io | 62 |
| github.com | 49 |
gartner.com first and youtube.com third are the two lines a reader will repeat. Only two vendors' own sites are cited more than either: the engines quote the analyst and the video before they quote most vendors' pages.
Technically ready, rarely named
Site readiness is high across the set (median 83) and visibility is low (median 30). Sherlocks AI scores 100 on the thirteen technical checks and 14 on visibility; Causely scores 85 on the thirteen technical checks and 15 on visibility; Checkmk scores 87 on the thirteen technical checks and 22 on visibility.
The technical work is largely done in this category. What is missing is being named by the sources in the table above, which no change to a website produces on its own.
What the named ones have in common
Only what the answers show. The brands that take questions are cited from gartner.com and g2.com, have comparison pages of their own that the engines quote, and appear in the citations under their own domain. That is three things, and the report lists which of them a given company is missing.
Check your own
If you are one of the 20, your full report already exists: enter your website on the home page and it is handed back in under a minute. If you are not, the scan takes three minutes.
Methodology and data as of 2026-09-15. Updated when a new batch runs in the category.
About this data
How were the companies chosen?
They are companies in AIOps, observability and monitoring that we scanned in September 2026, either for a client or as prospects. It is not a random sample of the category and the page does not claim to be; it is every scan we ran, with nothing left out.
What does a question look like?
Derived from each company's own site: 'best AIOps platform for a hybrid cloud estate', 'X alternatives for mid-size SRE teams', 'X vs Dynatrace'. About fifteen per company across four buckets.
Can a company on this list see its full report?
Yes. Enter the website on the home page; a website scanned recently is handed its existing report rather than a new scan, and the PDF opens for a work email.
How often is this updated?
When we run a new batch in the category, and the date at the top changes when we do.
Where do you stand in your category?
Your website and region. The score appears in about three minutes; the PDF opens for a work email.